Volume 04, No. 02, Feb 2018
P
age
1
Designing Combo Recharge Plans for Customers using Apriori
Algorithm
Tanisha Ingale*, Shashank Damani**, Deepshikha***,
Aditya
Karmalkar****, Santosh Darade*****
Department Of Computer Engineering, Sinhgad Institute of Technology and Science, Pune-41
ABSTRACT:
Machine Learning has become an integral part of human research now a day. People are tending to select more automatic system rather than going with the manual handling. Data mining has the huge effect on business analysis as all business rely on their behaviour of customers. Mining the behaviour of customers can help the very existence of the company. For mining such kind of data, association rules are used. This algorithm helps in finding the itemsets that are used frequently. This paper has proposed the way to satisfy customers in telecommunication market. Knowing the customer’s recharge pattern can enhance their will to use the same service provider. By mining the recharge pattern of individual customer, this system should be able to provide a new combo of recharges which will indeed be less than the individual recharges. For mining such frequent itemsets, this paper has used Apriori
algorithm.
KEYWORD: Data Mining, internet based facilities, SMS, Association Rule Mining,
Clustering, Frequent Item-set, Apriori Algorithm, Churn Rate.
INTRODUCTION:
Volume 04, No. 02, Feb 2018
P
age
2
combo offers that will bundle all important services (like talk value, data, SMS, Value Added Services (VAS) etc.) in such a way to make most of the customers happy (with the offer) as well as maximize revenue per user from all services.
Figure 1 : Different Types of Recharges
Data mining is the technique of sorting through large data storage to identify patterns and establish relationships to solve problems using data analysis. Data mining is the process of analyzing large amounts of data stored in a data warehouse for useful information which makes use of artificial intelligence, neural networks, and advanced statistical tools to reveal trends, patterns and relationships. Data mining tools allow business market to predict future trends. Information generated through data mining is used for decision making. Data mining supports many different techniques for knowledge discovery and prediction such as classification, clustering, sequential pattern mining, association rule mining and analysis, sequence or path analysis. Data mining is mainly used in mathematics, cybernetics, genetics and marketing, business analysis, strategic decision making, financial forecasting, future sales prediction etc. Our proposed model uses association rule mining technique. This technique is used to show relationship between different items. By analyzing large dataset correlation among items can be analyzed. Market basket analysis is another name given to association rule mining technique.
Two important concepts of association rule mining are :-
1. Support :- It is the count of the itemset in the total number of transactions. The support for an association rule X->Y is the percentage of transactions in the database that contains X U Y.
If A->B
Support (A->B) = Tuples containing both A and B
____________________________________ Total number of tuples
2. Confidence :- It is the ratio of number of transactions that contain A U B to the number of transactions that contain A.
Confidence (A->B) = Tuples containing both A and B
Volume 04, No. 02, Feb 2018
P
age
3
Literature Survey:
Author: Giridhar Maji, Soumya Sen
Title: Data Warehouse Based Analysis on CDR to Retain and Acquire Customers by
Targeted Marketing.
Objective: In this paper a mechanism is developed to store CDR data in a suitable Data
Warehouse (DW) schema and analytically process these using OLAP tools to understand the prepaid customers usage, spending and propensity to marketing offers.
Author: Owczarczuk, Marcin
Title: Churn models for prepaid customers in the cellular telecommunication industry using
large data marts.
Objective: This paper studies usefulness of the popular data mining models to predict churn
of the clients of the Polish cellular telecommunication company.
Author: Dairo, Adeolu O.
Title: Customer base management in a prepaid mobile market: Usage risk and usage
opportunity model.
Objective: In this paper, Usage Risk and Usage Opportunity model was developed that can
be used to understand critical usage behavior within the customer base.
Author: Verbeke, Wouter, David Martens, and Bart Baesens.
Title: Social network analysis for customer churn prediction.
Objective: Social network analysis has been used to predict churn in using relational learning
algorithms to incorporate social network effects within a customer churn prediction setup.
Author: Gerpott, Torsten J., and Phil Meinert.
Title: Choosing a wrong mobile communication price plan: An empirical analysis of
predictors of the degree of tariff misfit among flat rate subscribers in Germany.
Objective: This paper presents the relationship between the degrees of tariff choice misfit
among mobile service subscribers with customer characteristics.
PROBLEM STATEMENT:
Motivation:
1. The introduction of Reliance Jio lead to the downfall of other telecom servies.
2. Due to increasing Churn rate, there is a need of a system which can maintain the
Volume 04, No. 02, Feb 2018
P
age
4
SYSTEM ARCHITECTURE:
Fig.2:System Architecture
OUR APPROACH:
1. Firstly, the users will register their account to the system. This will include their Name,
Age, Address, Registered Mobile Number, Identity Proof, Nationality etc. After the submit button, the identity of the user will be verified by sending OTP to their e-mail ID. The information of registered users will be stored in the database.
2. After this, the user will be logged in to the system by giving their credentials. This
credential will be tallied through the database.
3. After logging in, user will be redirected to their home screen where they need to log their
previous amount and scheme they used. Each time the user recharges their phone, their data will be stored in the database.
4. By having sufficient number of user’s monthly recharges, our next step is to extract their
interest as to which scheme they tend to satisfy.
5. Once their frequent recharges have been extracted and their recharging pattern has been
recognized, the combination of schemes will be suggested to them. This will satisfy the customer’s requirement.
CONCLUSION:
Volume 04, No. 02, Feb 2018
P
age
5
to generate combo data plans after analysis. Expectation is that these combo plans with cost price slightly less than the individual price combined will make most of the customers happy as well as telecom service provider can be sure to get full potential revenue from those customers.
REFERENCES:
i. Owczarczuk, Marcin. "Churn models for prepaid customers in the cellular
telecommunication industry using large data marts." Expert Systems with Applications 37, no. 6 (2010).
ii. Maji, G., Sen, S. "Retaining and Acquiring New Customers by Targeted Marketing:
Analysis using DW on CDR Data", IEEE 5th International Conference on Reliability, Infocom Technologies and Optimization (ICRITO 2016) Noida, India, Sept 2016.
iii. Dairo, Adeolu O. "Customer base management in a prepaid mobile market: Usage
risk and usage opportunity model." African Journal of Marketing Management 7, no. 3 (2015).
iv. Verbeke, Wouter, David Martens, and Bart Baesens. "Social network analysis for
customer churn prediction." Applied Soft Computing
v. Gerpott, Torsten J., and Phil Meinert. "Choosing a wrong mobile communication
price plan: An empirical analysis of predictors of the degree of tariff misfit among flat rate subscribers in Germany.“ Telematics and Informatics (2016).
vi. Backiel, Aimée, Bart Baesens, and GerdaClaeskens. "Predicting time-to churnof
prepaid mobile telephone customers using social network analysis." Journal of the Operational Research Society (2016).
vii. Hung, Shin-Yuan, David C. Yen, and Hsiu-Yu Wang. "Applying data mining to
telecom churn management." Expert Systems with Applications31, no. 3 (2006): 515-524.
